Variations in state welfare policies in the reform era may affect adolescents through two mechanisms: A competing labor market hypothesis posits that stringent state welfare policies may reduce adolescent employment; and a signaling hypothesis posits that stringent welfare policies may promote enrollment. To test these hypotheses, we use a dynamic joint model of adolescents' school enrollment and formal employment, separating state welfare policies from non-welfare state policies, state labor market conditions, and unobserved state characteristics. Longitudinal data from the NLSY97 on adolescents aged 14 to 18 and various state data sources over the period 1994-1999 support the competing labor market effect but not the signaling effect. In particular, lower-income dropouts suffer more severely from fewer labor market opportunities when state welfare policies are more stringent, which indicates that welfare reform may compromise work opportunities for lower-income dropouts.